Swa-perlanggaran fabrik menggunakan gabungan pepohon-r dan lengkungan

Abstract

In fabric modeling and visual display on computer, self-collision is a collision process of points on the surface that exist in the fabric model. Self-collision consists of two phases; firstly is to detect the existence of collision points and secondly is to determine the result whether they are collision or adjacent points. The fabric model is generated using spring-mass model, which takes into account the elastic and damping forces as the internal forces and the gravity as the external forces when the object is moving. The objective of this research is to generate a self-collision technique for fabric model by combining the R-tree technique and curvature method. The R-tree method is used to allow only a minimum number of traversing nodes in detecting collision among the potential collision points and the curvature method is used to determine the actual location of the collision points. The research result is the fabric model which has four fixed diagonal in a closed environment without other objects and together with three factors, that are spring constant, mass and damping. The test is conducted to make comparison between the model with self-collision and the model without self-collision. These two models are generated to evaluate the effectiveness of self-collision. Case studies are carried out to determine the limit for the three foregoing factors to avoid penetration. The test result indicates that the maximum value of spring constant, mass and damping is 15.0, 0.005 and 0.95 respectively which causes penetration on the fabric model that leads to the unstable and broken fabric model. The overall result has shown that the generation of the fabric model that incorporates with self-collision gives a better fabric model visual display.